Re: [gentoo-dev] Native vs Scripting language for portage speed concerns was -> Sets vs Meta ebuilds

[William L. Thomson Jr.]

On Mon, 10 Jul 2017 20:12:28 -0500
R0b0t1 <r030t1@×××××.com> wrote:

> On Mon, Jul 10, 2017 at 6:56 PM, William L. Thomson Jr.
> <wlt-ml@××××××.com> wrote:
>>
> > IMHO anyone complaining about time taking for dependency resolution
> > etc. They should spend THEIR time writing stuff in a real native
> > language for speed.
> >
> > The difference I see with jem[1] vs java-config, is ridiculous.
....
> > If aspects of portage were done in C or C++, or maybe even Go. There
> > would be substantial performance improvements. The existing python
> > code can remain. Python can load and call functions from C/C++ DSO.
> > And vice versa, calling Python code from C/C++. I would say C vs
> > C++ but that is up to others.
> >
>
> https://wiki.gentoo.org/wiki/Q_applets
>
> What you're suggesting is actually really hard, and the root of the
> problems tend to be graph traversals and path searches (for
> dependencies) not so much miscellaneous milliseconds spent in
> interpreter overhead.

I am aware in a way. Depends on how implemented. This has to hit
package.env files. But what you see below comes from a dependency list.
I have packages with even more deps.

jem --help

 Package Options:
  -d, --with-dependencies    Include package dependencies in
                                        --classpath and --library calls
  -p, --classpath=PACKAGE(s) Print entries in the environment classpath
                                        for these packages

jem -d -p eclipse-jdt-core-4.6
/usr/share/ant-core/lib/ant.jar:/usr/share/ant-core/lib/ant-bootstrap.jar:/usr/share/ant-core/lib/ant-launcher.jar:/usr/share/eclipse-core-contenttype-4.6/lib/eclipse-core-contenttype.jar:/usr/share/eclipse-core-filesystem-4.6/lib/eclipse-core-filesystem.jar:/usr/share/eclipse-core-jobs-4.6/lib/eclipse-core-jobs.jar:/usr/share/eclipse-core-resources-4.6/lib/eclipse-core-resources.jar:/usr/share/eclipse-core-runtime-4.6/lib/eclipse-core-runtime.jar:/usr/share/eclipse-equinox-app-4.6/lib/eclipse-equinox-app.jar:/usr/share/eclipse-equinox-common-4.6/lib/eclipse-equinox-common.jar:/usr/share/eclipse-equinox-preferences-4.6/lib/eclipse-equinox-preferences.jar:/usr/share/eclipse-equinox-registry-4.6/lib/eclipse-equinox-registry.jar:/usr/share/eclipse-osgi-4.6/lib/eclipse-osgi.jar:/usr/share/eclipse-text-4.6/lib/eclipse-text.jar:/usr/share/osgi-core-api-6/lib/osgi-core-api.jar:/usr/share/eclipse-core-expressions-4.6/lib/eclipse-core-expressions.jar:/usr/share/osgi-compendium-6/lib/osgi-compendium.jar:/usr/share/eclipse-osgi-services-4.6/lib/eclipse-osgi-services.jar:/usr/share/osgi-annotation/lib/osgi-annotation.jar:/usr/share/felix-resolver/lib/felix-resolver.jar:/usr/share/eclipse-core-commands-4.6/lib/eclipse-core-commands.jar:/usr/share/icu4j-59/lib/icu4j.jar:/usr/share/glassfish-persistence/lib/glassfish-persistence.jar:/usr/share/osgi-foundation/lib/org.osgi.foundation.jar:/usr/share/tomcat-servlet-api-4.0/lib/el-api.jar:/usr/share/tomcat-servlet-api-4.0/lib/jsp-api.jar:/usr/share/tomcat-servlet-api-4.0/lib/servlet-api.jar:/usr/share/eclipse-jdt-core-4.6/lib/eclipse-jdt-core.jar

jem
real    0m0.004s
user    0m0.000s
sys     0m0.005s

java-config
real    0m0.075s
user    0m0.057s
sys     0m0.017s

On further runs of java-config
real    0m0.121s
user    0m0.105s
sys     0m0.017s

jem never exceeds 0m0.005s for any real, user, or sys

> Then again, I suppose there will be people on computers slow enough
> where the latter does make a difference.

arm? Laptops? Even if not speed, resources. But when its used for many
packages called over and over. It can start to add up little by little.
We can all use more time.

-- 
William L. Thomson Jr.